3 ms·2-D Sparse Parallelism for Deep Learning Recommendation Model Training1 points by PaulHoule 1y ago